RSI Trading Strategy – Does it Work Best During Asia Trading Session?
Intraday forex market seasonality points to significantly different market conditions depending on trading session and time of day. How can we shift trading techniques to take advantage of such moves? This report looks at the simple Relative Strength Index (RSI) Trading strategy to take advantage of changing market conditions.
Relative Strength Index – Range Trading Strategy of Choice, but How Can it be Improved?
The Relative Strength Index is one that has featured prominently in past DailyFX Strategy reports, as its popularity and reasonably good track record makes it a good benchmark range trading strategy. In past articles we have discussed setting stops for the RSI Strategy and using volatility filters with the RSI with good results. In particular we noted that the RSI tends to do poorly during times of high market volatility. Thus it seems reasonable we may look to explore intraday trends in volatility and attempt to use similar filters to avoid poor conditions for the RSI trading strategies.
Intraday Seasonality – What is it and how can we use it?
Given that the forex market is open 24 hours a day, it is important to note key differences in three distinct trading sessions and use this information to our advantage.
As the charts show, volatility is most often highest through the overlap between the London trading session (approximately 03:00 – 11:00 Eastern Time) and the New York session (approximately 09:00 – 17:00 ET) for major currency pairs. If we know that a strategy is likely to underperform amidst the sharpest currency moves, then we should probably avoid trading through such times. It seems worthwhile to explore the concept of a time filter for the RSI strategy.
Shutting off RSI Strategy during most volatile times of day
When we discussed volatility filters for the RSI, we found that the strategy tended to underperform during the most active market conditions. Thus we will look to use the same concept on an intraday level and with the simplest rules possible from the start.
As a raw strategy the RSI has been rather underwhelming on a 15-minute EURUSD chart going back to 2001. Though it shows some periods of strong results, fairly frequent sharp declines mean that the equity curve slopes sharply downward.
Source: FXCM Strategy Trader.
Looking for periods of low volatility for RSI Strategy
Focusing solely on the Euro/US Dollar chart, we see that volatility tends to drop off noticeably in one key hour through every trading session. Namely, in the final hour of the New York session (16:00-17:00), halfway through the London trading period, and towards the end of the Tokyo trading day. It is likewise clear that the strongest volatility tends to occur through late London and early New York trading hours—potentially warning against trading any strategies especially vulnerable to sharp currency moves.
RSI Trading Strategy with Time Filter
Using Strategy Trader, we can import a readily available RSI Trading strategy. But we’re going to go one step further and establish a slightly modified version.
Entry Rule: When the 14-period RSI crosses above 30, buy at market on the open of the next bar. When RSI crosses below 70, sell at market on the open of the next bar.
Filter: Strategy cannot enter trades between the start hour (startTime) and end hour (endTime). Yet it will not close any open trades at end hour and will hold them open until the reverse signal is triggered. (More on that topic later)
Stop Loss: None by default
Take Profit: None by default
Exit Rule: Strategy will exit a trade and flip direction when the opposite signal is triggered.
Backtesting our Time Filter For the Relative Strength Index Trading Strategy
In order to test the validity of this filter, we of course need to establish which times we would like to start trading and stop trading altogether. Given what we saw with the Euro/US Dollar, it seems logical to try limiting the system to the least-volatile hours of the day—punctuated by volatility low-points in the late New York session and midway through Tokyo trading. Thus our ‘StartTime’ variable will be set at 16:00 and ‘EndTime’ at 00:00 Eastern time. Below is the resulting equity curve.
Certainly this is an improvement over the base case of steady losses. Yet the fact that the equity curve has done little other than decline in the past 2 or so years does not bode well for future prospects, and indeed we may need to explore further options as far as our start and end times are concerned. Thus we turn to optimization.
Optimizing against two variables
Using Strategy Trader we will optimize against two variables in order to maximize theoretical profits. When we optimize we always want to find the best hypothetical risk-adjusted returns. And though we will keep the limitations of hypothetical trading strongly in mind, looking at what has worked in the past gives us a better sense of what is more likely to work in the future. We will optimize to maximize “Return on Account”, or the amount by which final strategy equity exceeds the minimum account size needed to run the strategy through the testing period. Minimum account size is determined by the maximum theoretical drawdown of the particular strategy. Thus our “Return on Account” variable will give us Final Profit/Loss over Maximum Drawdown.
Three-Dimensional Optimization Results Chart of Time Filter on EURUSD RSI Trading Strategy
Data source: FXCM Strategy Trader. Graph source: R-Project, RGL
It is admittedly difficult to properly display a 3D chart on a 2D medium, but the optimization results chart is quite revealing. Our best “Return on Account” percentages seem to cluster around the same ‘EndTime’ values, while the results on changing ‘StartTime’ values seem much more varied.
More concretely, our strategy tends to have the best results for the EURUSD when the ‘EndTime’ for our filter is between the hours of 05:00 to 07:00 Eastern Time. The best theoretical risk-adjusted returns will likewise come when our ‘StartTime’ is between the hours of 14:00 and 18:00 hours. What is our absolute best hypothetical backtest result?
Euro/US Dollar RSI Strategy Restricted to Trade between Hours of 14:00 and 06:00 Eastern Time
Source: FXCM Strategy Trader.
Are we done? No. We have established that this strategy has theoretically worked very well on the Euro/US Dollar through the past, but this is hardly a guarantee of future results. We are always in search of the most robust and stable result that is likely to work well in the future. One of the ways I have personally checked optimization results is to make sure they are consistent across currency pairs.
At this point it serves to mention that all currency pairs are not created alike, and this is especially true given that traders across the world will tend to trade more heavily in their regional currencies. Thus the Japanese Yen will see higher volatility during Asia hours than will the Euro or British Pound. It subsequently makes sense to compare currencies of the same region against each other when running robustness checks. And though the chart below does not show an exhaustive list of all possible time combinations, I chose several time limits that tended to work best across key currency pairs.
Given that the EURUSD and USDCHF have historically been very highly correlated, it should come of little surprise that the 14:00-06:00 time filter works very well for both. And while it does not work quite as well on the GBPUSD pair, it is still a vast improvement over the base RSI strategy and theoretically produces positive final equity.
We likewise note that the time frames restricted to times of much lower volatility did not perform nearly as well on these currencies as has the fairly wide 14:00-06:00 stretch. The RSI strategy tends to lose during times of especially strong price moves, but it does need some volatility to generate trades. That is perhaps one reason why our top time frame includes some fairly volatile periods for each the EURUSD, USDCHF, and GBPUSD currency pairs. What of Asia-centric currencies?
Unfortunately our time filter does not work nearly as well on the USDJPY, GBPJPY, AUDUSD, and NZDUSD—not producing any one positive equity curve across the same testing period. And though the 20:00-03:00 stretch shows some promise in recent years, it does not look nearly as impressive as our top equity curve in European currency pairs. A closer look at the equivalent optimization graph for the US Dollar/Japanese Yen pair highlights a key reason this is the case.
Three-Dimensional Optimization Results Chart of Time Filter on USDJPY RSI Trading Strategy
Data source: FXCM Strategy Trader. Graph source: R-Project, RGL
Due to the JPY’s relatively high volatility through Asian trading hours, it seems that time-filtering the RSI system to specific hours has little effect. And though the AUDUSD and NZDUSD see slightly better results, it is not enough to produce an overall positive equity curve. Our time-filtered RSI system seems to be better suited to European and North American currency pairs.
Backtesting Time Filters on RSI Strategy – Showing Promise
Initial results on our time filters show promise with the RSI Trading Strategy on the EURUSD, GBPUSD, and USDCHF pairs. Said data suggests there is further work to be done, and we have the makings of a potential winning strategy based on hypothetical backtests. Yet the current logic exposes us to entirely too much risk during the most volatile trading hours of the day. That is, the rules dictate that we cannot open or close trades outside of our trading periods—allowing for potentially disastrous losses.
The next installment of our Forex Strategy Corner will take a closer look at said strategy and attempt to make it a more suitable to real trading. That said, we could easily see such time filters working on a broad range of range trading systems—not limited to our go-to RSI benchmark.
If you would like to suggest ideas for this topic or any other forex strategy you would like to see in this series, feel free to e-mail author David Rodríguez at email@example.com. To be added to this author’s distribution list, e-mail with subject line “distribution list”
View previous articles in this series:
Written by David Rodríguez, Quantitative Strategist for DailyFX.com
DailyFX provides forex news and technical analysis on the trends that influence the global currency markets.